by Michael Pettis, China Financial Markets
On Tuesday the National Bureau of Statistics released China’s 2014 GDP growth numbers and reported growth consistent with what the government has been widely promoting as the “new normal”.
According to the preliminary estimation, the gross domestic product (GDP) of China was 63,646.3 billion yuan in 2014, an increase of 7.4 percent at comparable prices. Specifically, the year-on-year growth of the first quarter was 7.4 percent, the second quarter 7.5 percent, the third quarter 7.3 percent, and the fourth quarter 7.3 percent.
As nearly every news article has pointed out, GDP growth of 7.4% slightly exceeded consensus expectations of around 7.3%, setting off a flutter in the Shanghai Stock Exchange that reversed nearly a quarter of Monday’s disastrous drop of almost 8%. But although it exceeded expectations 2014 still turned in the lowest reported GDP growth since 1990, presenting only the second time since growth targeting began in 1985 that the reported number came in below the official target (the first time was in 1989). We will undoubtedly be swamped in the coming days with analyses of the implications, but I read the data as telling us more about the state of politics than about the economic health of the country.
Over the medium term I have little doubt that growth will continue to slow, and that at best we have only completed about one third of the journey from peak GDP growth to the trough. As long as it has debt capacity Beijing, or indeed any other government, can pretty much get as much growth as it wants, in China’s case simply by making banks fund local government infrastructure spending.
Most economic analyses of the Chinese economy tend to base their forecasts on the sequence and pace of economic reforms aimed at rebalancing the economy, and on the impact these reforms are likely to have on productivity growth. It may seem contrarian, then, that I forecast Chinese near term growth largely in terms of balance sheet constraints. I am not implying that the reforms do not matter to the Chinese economy. The extent to which the reforms Beijing proposes to implement reduce legal and institutional distortions in business efficiency, eliminate implicit subsidies for non-productive behaviour, reorient incentives in the capital allocation process, undermine the ability of powerful groups to extract rent, and otherwise liberalise the economy, will unquestionably affect China’s long-term growth prospects.
But I expect that any significant impact of these reforms on short-term growth will largely be the consequence of two things. The first is how reforms will affect the amount, structure, and growth of credit. The second is how successfully Beijing can create sustainable sources of demand that do not force up the debt burden — the most obvious being to increase the household income share of GDP and to increase the share of credit allocated to small and medium enterprises relative to SOEs.
To put it a little abstractly, and using a corporate finance model to understand macroeconomics, I would say that most economists believe that China’s growth in the near term is a function of changes in the way the asset side of the economy is managed. If Beijing can implement reforms that are aimed at making workers and businesses utilize assets more productively, then productivity will rise and, with it, GDP.
This sounds reasonable, even almost true by definition, but in fact it is an incomplete explanation of what drives growth. In corporate finance theory we understand that although growth can often or even usually be explained as a direct consequence of how productively assets are managed, it is not always the case that policies or exogenous variables that normally change the productivity of operations will have the expected impact on productivity growth. When debt levels are low or when the liability structure of an economic entity is stable, then it is indeed the case that growth is largely an asset-side affair. In that case for GDP growth to improve (or for operating earnings to rise), managers should focus on policies aimed at improving productivity.
But when debt levels are high enough to affect credibility, or when liabilities are structured in ways that distort incentives or magnify exogenous shocks, growth can be as much a consequence of changes in the liability side of an economy as it is on changes in the asset side. At the extreme, for example when a company or a country has a debt burden that might be considered “crisis-level”, almost all growth, or lack of growth, is a consequence of changes in the liability structure. For a country facing a debt crisis, for example, policymakers may work ferociously on implementing productivity-enhancing reforms aimed at helping the country “grow” its way out of the debt crisis, but none of these reforms will succeed.
When liabilities constrain assets
That both orthodox economic theory and government policy-making ignore the way liability structure can overwhelm the impact of asset-side management is surprising given how strong the historical confirmation. There is a long history of countries either facing debt crises or struggling with dangerous debt burdens — including many countries today both in the developed world and the developing world — in which policymakers have promised to implement dramatic policies that will improve productivity and return the economy to “normalcy”. But just as today growth stubbornly stagnates or decelerates in Europe, Japan, China, and a number of over-indebted countries, it is hard to find a single case in modern history in which a country struggling with debt has been able to reform and grow its way out of its debt burden until there has been explicit or implicit debt forgiveness. It is no accident that growth in Japan, China and Europe keep disappointing analysts, and on Tuesday the IMF yet again cut its global growth forecast by 0.3% — to 3.5% and 3.7% in 2015 and 2016. It will almost certainly continue to cut it over the next few years.
In the book I plan to write this year I hope to explore the conditions under which the structure of liabilities matter to growth, and to show how sometimes it is even the only factor that determines growth rates. It is not just as a constraint that a country’s liability structure affects growth, however. There are times when it can actually reinforce growth. There are in fact many ways in which a country’s balance sheet can significantly affect growth rates, both during growth acceleration and growth deceleration, and China demonstrates just such a case.
In fact as far as I can tell, in every case in modern history of very rapid, investment-driven growth, at least part of the growth was caused by self-reinforcing credit structures embedded in the balance sheet. One way is by encouraging additional investment to expand manufacturing and infrastructure capacity. Rapid growth raises expectations about future growth, making it easy to fund projects that expand capacity even further, and these projects themselves result in faster growth, which then justifies even higher growth expectations. Another way is by improving credit perceptions. When loans are backed by assets, rapid growth increases the value of these assets, so that the riskiness of the existing loan portfolio seems to decline, allowing the lender to increase his risky loans and the borrower to increase his purchase of assets, which of course puts further upward pressure on asset values.
There is nothing surprising about either process — we all understand how it works. But what we sometimes forget is that when this happens economic activity can easily exceed the increase in real economic value-creation, and more importantly, the same balance sheet structure can cause growth to decelerate far faster than we had expected during the subsequent adjustment period. The balance sheet causes growth to be higher than it would have otherwise been during the growth phase and slower when growth begins to decelerate. It is not just coincidence that nearly every case in modern history of a growth miracle has been followed by a brutally and unexpectedly difficult adjustment. The same balance sheet that turned healthy growth into astonishing growth turned a slowdown into a collapse.
My 2001 book, The Volatility Machine, was about the history and structure of financial crises in developing countries, and in the book I discuss some of these balance sheet structures that exacerbate both accelerating and decelerating growth. In this essay I want to discuss concrete examples of such structures and show how they impact growth. In the book I distinguish between “inverted” and “hedged” balance sheets, and it is worth explaining the distinction. A hedged balance sheet is simply one that is structured to minimise the overall volatility of the economic entity, whether it is a business or a country.
When the balance sheet is fully hedged, the only thing that changes its overall value is a real increase in productivity. Any exogenous shock that affects the value of liabilities and assets, or that affects income and expenditure, will have opposite effects on the various assts and liabilities, and together these will add to zero. Of course the closer an economic entity is to having a perfectly hedged balance sheet, the lower the cost of capital, the lower the rate at which expected earnings or growth is discounted over time, the easier it is for businesses to maximise operating earnings without worrying about unexpected shocks, and the longer the time horizon available for both policymakers and businesses in planning.
An inverted balance sheet is the opposite of a “hedged” balance sheet, and involves liabilities whose values are inversely correlated with asset values. These embed a kind of pro-cyclical mechanism that reinforces external shocks by automatically causing values or behavior to change in ways that exacerbate the impact of the shock. When asset values rise, in other words, the value of liabilities falls (or, to put it differently, the cost of the liabilities rise), and vice versa.
Balance sheet inversion
A business or country with an inverted balance sheet benefits doubly in good times as its assets, or its earnings, rise in value and its liabilities, or its financial expenses, fall. The process is often self-reinforcing, especially when the inverted entity is a country, in which case the economy can be described as being in a virtuous circle. When Brazil began to reform its economy and instituted a new currency regime in 1994, for example, one of its greatest vulnerabilities was its extremely high fiscal deficit, more than 100% of which was explained by debt servicing costs. Most Brazilian government debt was of less than six months maturity, and nearly all of it matured within one year (short-term debt is extremely inverted). As Brazilian reforms associated with the 1994 currency regime increased overall confidence, short-term interest rates declined, and within months the fiscal deficit followed suit. This caused confidence to rise sharply, and interest rates to fall further. In Brazil interest rates fell steadily from well over 50% in the early 1994-95 to around 20% by the summer of 1998.
Of course in bad times the opposite happens – the value of assets fall while the value of liabilities rise, and the virtuous circle quickly becomes a vicious circle. Financial distress costs are not linear, and so it is not surprising that conditions usually deteriorate much more quickly than they improve. When the Russian crisis in 1998 shook confidence in emerging markets, Brazilian interest rates suddenly began rising, which caused the fiscal deficit to shoot up and so undermined confidence further, locking the country into an extremely vicious circle that took interest rates back to over 40% within two or three months. In January of the following year Brazil was forced into a currency crisis.
Inverted balance sheets, in other words, automatically exacerbate both good times and bad. Among other things this often leads to confusion about the sources of growth and value creation and the quality of management. When an economy is doing well the short-term gains for the economy that are simply a consequence of balance sheet inversion are often treated in the same way as ordinary productivity gains caused by better management when we try to judge the effectiveness of the underlying economic policies. In reality, however, they are just forms of speculative profits.
This may seem a surprising statement, but in the Brazilian case described above, for example, while part of the decline in the fiscal deficit before the summer of 1998 can be explained by better policies, at least part of the decline came about simply because of the very short debt maturities. If the Brazilian government had funded itself with longer-term debt — which would have been much more appropriate and far less risky — the fiscal deficit would not have declined nearly as quickly as it did. Part of the improvement in the fiscal deficit, in other words, was simply the consequence of what was effectively a speculative bet on declining rates, and did not reflect better fiscal policies. One unfortunate consequence, however, was that analysts and policymakers overvalued the quality and impact of government policies.
Of course when conditions turn, inverted balance sheets also provide short-term losses, although, perhaps not surprisingly, managers or policymakers almost always recognise the component of “bad luck” in their weakened performance. In 1998 I had many conversations with Brazilian central bankers, including the president of the central bank, Gustavo Franco, about taking advantage of high confidence in Brazil to borrow long-term at rates actually below the then-current 1-year rate of 20% (we were prepared to raise $1 billion of five-year money at 19%). The central bank decided against doing so at least in part because they were confident that the market was responding mainly to the quality of their monetary policies, and that as they were determined to maintain these policies, they felt it did not make sense to extend maturities until interest rates had dropped by far more. My carefully worded suggestions that at least part of their success was the result of an implicitly speculative balance sheet, and that it might make sense to reduce that risk, were not well-received.
Of course when interest rates shot up, no one doubted the role of balance sheet structures in the subsequent crisis. My point here is not a cynical one about the vanity of policymakers. It is that when a country responds very positively to policy reforms it is genuinely difficult for most economists to distinguish between growth caused by the reforms and growth caused by the self-reinforcing nature of inverted balance sheets, and the more highly inverted a country’s balance sheet, the more dramatically will good policies seem to be rewarded. Over the long term, however, because the same virtuous circle can become an equally powerful vicious circle, inverted balance sheets always automatically increase financial distress costs because for any level of debt it increases the probability of default.
Along with short-term debt as described above, external currency debt is a very typical kind of inverted borrowing because when the borrowing country is growing rapidly, the tendency is for its currency to appreciate in real terms, and this reduces its debt servicing costs as interest and principle has to be repaid in cheaper foreign currency. Of course the opposite happens when the economy stagnates, and in a crisis a rapid depreciation of the currency can cause debt-servicing costs to soar exactly when it is hardest to repay the debt. The self-reinforcing combination of rapid GDP growth in the 1990s, reinforced by rapidly rising external debt, set the stage for the Asian Crisis of 1997, during which Asian borrowers were devastated as high levels of external debt caused growth to slow and currencies to weaken, both of which caused the debt burden to soar even faster.
Other types of inverted liabilities can include inventory financing, floating-rate debt, asset-based lending, margin lending, wide-spread use of derivatives, commodity financing, and real estate leverage (in countries in which the real estate sector has a major impact on GDP growth). High concentrations of debt in important sectors of the economy, even when in the aggregate debt levels are low, can also be important (and typical) forms of balance sheet inversion.
Developing countries historically have been very prone to creating inverted balance sheets during their growth phases, which is an important reason for their much greater economic volatility. This may simply be because often the least risky way of lending involves pushing the risk onto the borrower, for example by keeping maturities very short, or by denominating the debt in a more credible foreign currency. Because many developing countries are capital constrained, they often have no choice but to borrow in risky ways. In a 1999 paper, Barry Eichengreen and Ricardo Hausmann referred to this type of borrowing, “in which the domestic currency cannot be used to borrow abroad or to borrow long term even domestically”, as “original sin”.
Not included in the concept of original sin, but closely related, is the historical tendency of risk appetite to be highly correlated across the global economy. Foreign and local investors are most willing to lend to a risky developing country at a time when the whole world is benefitting from the easy availability of risk capital, and global growth is consequently high. Of course foreign capital dries up and local flight capital expands just as the world slows down and the economy begins to stagnate.
But there are other reasons developing countries typically build up inverted balance sheets. One reason may have to do with the ability of very powerful elites, in countries with limited separation of powers, low government accountability, and low transparency, to arrange that profits are privatized and losses are socialized. In that case it makes sense to maximize volatility, and inverted balance sheets do just that. Another reason may be the economic importance of commodity extraction to growth in many developing countries, and the tendency for capital to be available only when commodity prices are high and rising.
The confusion about whether rapid growth during reform periods has been driven more by virtuous circles or by virtuous policymaking of course also reinforces the tendency to increase inversion, especially in the late stages of a growth period when the economy reaches the limits of the growth model and begins naturally to slow. If at least part of the growth is the consequence of virtuous circles, as it usually is especially in a heavily credit-dependent economy, balance sheet inversion can be a bad short-term trading strategy because it increases the costs of an economic slowdown. If the growth is the consequence mainly of virtuous policymaking, as it is always believed to be, balance sheet inversion is a good short-term trading strategy because virtuous policymakers presumably will continue to put into place virtuous policies.
Whatever the source of growth, the already high economic volatility typical of developing countries is exacerbated by balance sheet structures that magnify this volatility. One consequence is that we are continually surprised by much more rapid growth than expected during good times and very rapid and unexpected economic deterioration just as things start to go bad.
We often see developing countries in the late, strained stage of a growth miracle rapidly build up inverted balance sheets even more quickly than earlier in the growth cycle. I am not sure why this is so often the case, but it could be that after many years of growth, reinforced by inverted balance sheets, economic agents become convinced that recent trends are permanent. They assume, for example, that interest rates must always drop, or that the currency always appreciates, or that real estate prices always rise, or that demand always catches up with capacity, so that it makes sense to bet that the future will look like the past.
There is also a natural sorting mechanism in a rapidly growing volatile economy in which business managers who tend, for whatever reason (including the ability of powerful vested interests to create asymmetrical distributions of profits and losses) to take on too much risk will systematically outperform and take market share from more prudent business managers. Anyone who is familiar with Hyman Minsky’s explanation of how attempts by regulators to reduce risk in the financial system will cause bankers to engage in riskier behaviour fully understands the mechanism.
There are several points that I think are useful when we think about how sensitivity to balance sheet structures might help us in forecasting growth, especially in countries that have a lot of debt and other institutional distortions:
All balance sheets are not the same. Liabilities can be structured in such a way that the performance of the asset side (or of operations) can be significantly affected. One of the ways in which this can happen is when liabilities exacerbate or reinforce operations or changes in the value of assets.These kinds of balance sheets are inverted, and they can embed highly pro-cyclical mechanisms into an economy.
There are many forms of inverted balance sheet structures, some of which are very easy to identify (external currency debt, margin financing, short-term debt) and others much more difficult to identify (an economy’s over-reliance on any single agent or industrial sector can create a kind of balance sheet inversion, for example, that is difficult to explain). The key consideration is when factors or policies that change underlying productivity are correlated, causally or not, with other parts of the balance sheet that affect the economy’s overall performance in the same direction.
Highly inverted economies are more likely to experience periods of exceptionally high growth or exceptionally deep stagnation, and the latter almost always follow the former. As far as I can tell, most growth miracles in modern history are at least partly the result of highly inverted balance sheets. This probably why they often seemed to grow far faster than anyone originally thought possible, and why most growth miracle economies subsequently experienced unexpectedly difficult adjustments.
It is often difficult to tell the difference between growth caused by fundamental changes in productivity and growth caused by pro-cyclical balance sheet structures — i.e. between virtuous polices and virtuous circles. This often causes analysts to overvalue the quality of policymaking or the underlying economic fundamentals during a period of rapid growth. As an aside, in my experience on Wall Street I can say that it can be very difficult to explain balance sheet inversion to the policymakers that preside over very rapidly growing economies, and it is never a good marketing strategy for a banker.
In the late stages of a period of rapid growth, as the economy is beginning to slow as it adjusts from the imbalances generated during the growth period, it seems to me that there is a systematic tendency to increase balance sheet inversion as a way of maintaining growth or of slowing the deceleration process.
Inventory can increase volatility
The last point of course is especially important in the Chinese context. The Chinese economy is clearly slowing, and debt is clearly rising. There is increasing evidence of highly inverted balance sheet structures within the Chinese economy, but I do not know if this is because balance sheet inversion is increasing or because a slowing economy causes pressure to build within the financial sector, and this makes visible risky structures that had been in place all along.
At any rate, it makes sense both for policymakers and for investors to try to get some sense of the extent of inversion in the economy. Most economists now expect that China’s economy will continue slowing, with most economists considering an eventual decline in GDP growth to 6% as the lower limit. Others, myself included, expect growth to slow much more than that. Of course the more inverted the Chinese balance sheet, the more any fundamental slowdown will be exacerbated by automatic changes in the country’s balance sheet.
This makes it very useful to get a sense of how balance sheet inversion can occur in China. Last Tuesday I saw three articles, two on Bloomsberg and one in the Financial Times that struck me as interesting examples of different kinds of balance sheet inversions. The first article reported that China was, according to Bloomberg, importing record amounts of crude oil as prices collapsed:
China’s crude imports surged to a record in December after a buying spree in Singapore by a state-owned trader and as the government in Beijing accelerated stockpiling amid the collapse in global oil prices.
…Chinese demand is shoring up the global oil market as the country expands emergency stockpiles amid crude’s slump to the lowest level in more than five years. The Asian nation’s consumption is forecast to climb by 5 percent in 2015, while the government is set to hoard about 7 million tons of crude in strategic reserves by the middle of this year, predicts ICIS-C1 Energy, a Shanghai-based commodities researcher.
Stockpiling oil in this case has a complex relationship within the balance sheet. On the one had it can be described as a kind of hedge. China is naturally short oil because it is a net importer. In that sense China benefits when the price of oil declines, and suffers when the price of oil rises.
Stockpiling oil, then, is a way of hedging. If oil prices continue to drop, China will lose value on its inventory position, but because the Chinese economy will be better off anyway, the losses China suffers from its stockpile simply reduce the overall benefit to China of lower prices. Meanwhile if oil prices should rise, China will suffer in the same way that all energy-importing countries suffer, but it will profit from its stockpile, and this profit will reduce the total loss. In that sense oil stockpiles reduce overall volatility in the Chinese economy, just as they do for any country that is a net importer of energy.
If China were a small country whose economic performance was largely uncorrelated with the economic performance of the world, this would be the end of the story. But China is not. Given how important the external sector is to China’s economy, growth in China is likely to be highly correlated with growth in the global economy.
This changes the picture. When the world is doing poorly, it is likely that oil prices will decline further and that China’s economy will do worse than expected. In that case, the more China stockpiles oil, the greater its losses. Oil prices in other words can be positively correlated with China’s economic performance, and stockpiling oil actually increases volatility because China profits on the stockpile when growth is higher than expected, and loses when it is lower than expected.
The second article, also in Bloomberg, told a similar story about iron ore:
Iron ore imports by China rebounded to an all-time high last month, capping record annual purchases, as slumping prices boosted demand for overseas supplies in the biggest user and some local mines were shuttered over winter.
China is by far the world’s largest consumer of iron ore, taking up very recently as much as 60% of all the iron ore produced in the world. What drives China’s voracious demand for iron, of course, is its extraordinarily high investment growth rate. For nearly five years I have warned that because I expected Chinese growth rates to drop significantly, I also expected the price of iron ore to collapse (and I have always added that in this context the word “collapse” was wholly appropriate). Clearly this has happened. There is a very high positive correlation between Chinese GDP growth and the price of iron ore, and so iron and steel inventory necessarily increases balance sheet inversion. If China slows further, it will take additional losses on its inventory as iron ore proies drop further. If Chinese growth picks up, China benefits from its stockpiling strategy.
Stockpiling iron ore might seem like a good idea for China if iron ore is so cheap that its price can no longer decline. In that case stockpiling iron or steel creates a hugely convex trade that more than compensates for the additional volatility that it adds to the Chinese economy. There are many investors, especially in China, who believe that iron ore prices have fallen so dramatically in the past two years that we have effectively reached the point at which the downside is minimal compared to the upside.
What to watch for
This may be true, but I think we have to be very skeptical about such arguments. Iron ore currently trades in the mid $60s, roughly one third of its peak in the $190s in late 2010 and early 2011, if I remember correctly. Many analysts believe that this decline has been so dramatic and astonishing that it cannot possibly continue. I disagree. At the turn of the century I think iron ore traded below $20, and it seemed at the time that prices could only decline even further. Current prices, in other words, only seem astonishingly low compared to their peak prices, which were driven by a surge in demand from China that was completely unprecedented in history. By historical standards, iron ore is not cheap at all. I have been arguing for years that a collapse in iron ore prices was inevitable, and iron would actually drop below $50 by 2016-17, perhaps even to $30-40 before the end of the decade, and I see no reason to assume that we are anywhere near the bottom.
Whether or not I am right about iron ore, the main point is that hard commodity prices have been driven to historically high levels largely because of China’s disproportionate share of global demand, with both prices and Chinese demand beginning to surge in 2003-04. Prices are highly positively correlated with Chinese growth, in other words, and stockpiling necessarily exacerbates both growth acceleration and deceleration.
Finally the third article, in the Financial Times, told what at first seemed to be a very different story:
Local governments in some of China’s smallest cities are snapping up an increasing amount of their own land at auctions, in a destructive cycle designed to prop up property prices but which is ravaging their own finances.
Local government financing vehicles in at least one wealthy province, Jiangsu, which borders Shanghai, accounted for more land purchases than property developers did in 2013 — the last year for which data were available — according to research collated by Deutsche Bank. The data signal that already cash-strapped local governments are switching money from one pocket to another rather than booking real sales.
Clearly it is extremely risky for local governments, who are highly dependent on land prices for their revenues, to increase their exposure to land prices by buying up land at auctions. This is an obvious case of balance sheet inversion at the local government level. Some economist might argue that while it may increase risk at the local level, it does not do so at the national level. It simply represents a transfer of wealth from one group of economic agents to another. If real estate prices fall, for example, local governments will be even worse off than ever, but property developers will be better off because they are less exposed than they otherwise would have been.
There are at least two reasons why this may be totally mistaken. The first reason is that by propping up real estate prices local government may be helping powerful local interests who only want to sell their real estate in order to fund disinvestment or flight capital. The second, and far more important, reason has to do with the fact that financial distress costs are concave, not linear. If there is a transfer of wealth from one indebted entity to another, the latter benefits at the former’s expense. But the reduction of financial distress costs for the latter must necessarily be less than the increase for the former. Taken together, there must be a net increase in financial distress costs for China and a net increase in volatility within China. This is not the place to explain exactly why this must happen (I will do so in my upcoming book), but if it were not true, then it would not be the case that a country could suffer from excessive domestic debt.
My main point is that orthodox economists have traditionally ignored the impact of balance sheet structure on rapid growth, but liability structures can explain both very rapid growth and very rapid growth deceleration. It is unclear to what extent balance sheet inversion explains part of the Chinese growth miracle of the past decade, but it would be unreasonable to discount its impact altogether, and I suspect it’s impact may actually be quite high. To the extent that it has boosted underlying growth in the past, for exactly the same reason it must depress underlying growth in the future.
What is more, because we are in the late stages of China’s growth miracle, we should recognise that historical precedents suggest that balance sheet inversion will have increased in the past few years, and may continue to do so for the next few years, which implies that a greater share of growth than ever is explained not by fundamental improvements in the underlying economy but rather by what are effectively speculative bets embedded into the national balance sheet. Besides commodity stockpiling and real estate purchases by local governments, we have clearly seen an increase in speculative financial transactions by large Chinese companies (the so called “arbitrage”, for example, in which SPEs have borrowed money in the Hong Kong markets and lent the money domestically to pick up the interest rate differential as well as any currency appreciation), which is the Chinese version of what in the late stages of the Japanese growth bubble of the 1980s was referred to as zaitech.
We have also seen growth in external financing, which is the classic form of inverted debt for developing countries. The main thing to watch for, I think, is one of the most dangerous kinds of balance sheet inversion, and is especially common when growth has been driven by leverage, and that is the tendency for borrowers to respond to credit and liquidity strains by effectively doubling up the bet and shortening maturities. I don’t know if this is happening to any worrying extent, but when we start to see a dramatic shortening of real maturities, it should be a warning signal.